Trains large language models (2B-462B parameters) using NVIDIA Megatron-Core with advanced parallelism strategies. Use when training models >1B parameters, need maximum GPU efficiency (47% MFU on H100
Install
The skill file stays in the publisher's repository.
Paste the copied command in a terminal. It installs this SKILL.md into that agent.
npx skills add Orchestra-Research/AI-Research-SKILLs --skill megatron-core -a cursor -y
Trains large language models (2B-462B parameters) using NVIDIA Megatron-Core with advanced parallelism strategies. Use when training models >1B parameters, need maximum GPU efficiency (47% MFU on H100), or require tensor/pipeline/sequence/context/expert parallelism. Production-ready framework used for Nemotron, LLaMA, DeepSeek.
name: training-llms-megatron
description: Trains large language models (2B-462B parameters) using NVIDIA Megatron-Core with advanced parallelism strategies. Use when training models >1B parameters, need maximum GPU efficiency (47% MFU on H100), or require tensor/pipeline/sequence/context/expert parallelism. Production-ready framework used for Nemotron, LLaMA, DeepSeek.
version: 1.0.0
author: Orchestra Research
license: MIT
tags: [Megatron-Core, Large-Scale Training, NVIDIA, Tensor Parallelism, Pipeline Parallelism, Model Parallelism, H100, Distributed Training, Production]
dependencies: [megatron-core, torch, apex, transformer-engine]
Megatron-Core - Large-Scale LLM Training
Quick start
Megatron-Core trains LLMs from 2B to 462B parameters with up to 47% Model FLOP Utilization on H100 GPUs through advanced parallelism strategies.
Installation:
# Docker (recommended)
docker run --gpus all -it --rm nvcr.io/nvidia/pytorch:25.04-py3
# Or pip
pip install megatron-core
Simple distributed training:
# Train with 2 GPUs using data parallelism
torchrun --nproc_per_node=2 examples/run_simple_mcore_train_loop.py
# Or LLaMA-3 8B training
./examples/llama/train_llama3_8b_fp8.sh
Common workflows
Workflow 1: Train LLaMA-style model with 3D parallelism
# Start with 1, increase until OOM
for MBS in 1 2 4 8; do
echo "Testing micro-batch-size=$MBS"
torchrun ... --micro-batch-size $MBS
done
Typical values:
7B model: 4-8
70B model: 1-2
405B model: 1
Step 4: Tune parallelism degrees
Rules of thumb:
Tensor Parallel: Use ≤8 (limited by NVLink within node)
Pipeline Parallel: Use for >70B models
Context Parallel: Use for sequences >8K tokens
Data Parallel: Fill remaining GPUs
Download 08-distributed-training/megatron-core/SKILL.md from the publisher, or copy the path and open the repository. PlugMCP lists the skill and does not host the file.
Where does training-llms-megatron live?
The publisher keeps it at https://github.com/Orchestra-Research/AI-Research-SKILLs.
Which agents can use training-llms-megatron?
Claude, Cursor, and Codex can read a SKILL.md you add from the publisher's repository.